A stochastic cost–benefit analysis framework for allocating energy storage system in distribution network for load leveling

Abstract

Increasing peak demand, retirement of conventional generation capacity, and high cost of constructing new generation capacity and network reinforcements are the challenges faced by utilities in urban cities. To address the above challenges, this paper proposes a stochastic cost–benefit analysis (CBA) framework, named CBA-LL, for allocating centralized energy storage system (ESS) in distribution network for load leveling. CBA-LL is built on a scenario-based stochastic optimization approach that integrates principal component analysis and k-means clustering for effective probabilistic load modeling. In CBA-LL, a genetic algorithm (GA) based optimization approach is utilized over a multi-year planning horizon for ESS sizing. Within the GA, the optimal operation of ESS for load leveling is determined using fmincon solver of MATLAB with the objective of minimizing load variance. The benefits of integrating ESS in terms of distribution system upgrade cost, distribution energy loss cost, arbitrage benefit, and generation plant upgrade cost are considered. In CBA for ESS allocation, the financial objective chosen is generally maximizing net savings (NS) (also known as net present value (NPV)). However, the limitation of NS objective function is that it does not consider the scale of investment required to achieve the objective. ROI and NS are two important and complimentary criteria in investment decision making. However, to the best of our knowledge, ROI has never been investigated in the literature as a planning objective function for sizing ESS. Hence, in this paper, we investigate maximizing ROI as a planning objective function and compare it with maximizing NS objective function. The proposed CBA-LL approach is implemented on a 33-bus distribution system, and two types of ESS technologies, namely, Li-ion and Vanadium Redox Flow (VRF) are compared. The simulation results demonstrate the complementary nature of NS and ROI objective functions. Furthermore, the simulation results highlight that ROI objective function may be more preferable when there is a lack of investment budget or high opportunity cost of investment. Our analysis suggests that the decision makers (DM) can investigate both NS and ROI objective functions for ESS sizing and then select a solution based on a more informed decision. The comparison of battery technologies demonstrate that the Li-ion battery is significantly superior to the VRF battery for the load leveling application.